This application proposes to develop predictive models which are aimed at explaining the release of chemical agents from the surface of particles into solvents which model physiological fluids found in the lung. Aromatic and polynuclear aromatic compounds and carbon particles of various origins will be used as examples of potential chemical agents and particles. These species have been selected since organic particulate matter has current environmental significance. The predictive models will be based of measurements of surface areas, surface heterogeneities, heats of adsorption and desorption, polarities, shapes and sizes. These parameters will be measured by volumetric, microscopic, gas solid chromatographic, liquid solid chromatographic and calorimetric procedures. Liquid solid chromatography is proposed as a novel method to study the release of agents in lung physiological fluids.